Heat transfer enhancement of carbon-nano fibers (CNF’s) attached on a wall surface within a micro-channel is investigated in this paper using a three-dimensional numerical method. Carbon-nanofibers, also known as graphite nanofibers, can be grown by catalytic decomposition of certain hydrocarbon at a metal surface such as iron, cobalt, nickel and some of their alloys. Typical sizes of CNF’s vary between 2 and 100 nm, with lengths ranging from 5 to 100 μm. Experimental research has shown that the presence of carbon-nano fibers grown on a surface of a fine metallic structure can enhance heat transfer by 50% [2]. These fibers influence the fluid flow, and enlarge the heat exchanging surface. The enhancement depends very much on the carbon-nano fibers density and on the structure of the carbon-nano fibers itself. This numerical study is giving directions in optimizing this new material. A random generation growth model has been developed to generate a stochastic structure of the CNF layer. Next to this a 3D Lattice Boltzmann model has been developed to simulate the heat transfer in a micro-channel flow with the surface covered with CNF’s. The 3D Lattice Boltzmann model has been verified on microchannel flow with heat transfer. Results of the conjugate heat transfer (including CNF’s at the wall) will be presented. The influence of carbon-nano fibers density and their structure on the heat transfer coefficient through the carbon-nano fibers layer is determined.